Flexible Microelectrode Arrays with Dural Regeneration for
Chronic Neural Recording
Tiffany
1*
Moy ,
Yan
2
Wong ,
Chiraag
1
Galaiya ,
3
Simon Archibald ,
Bijan
2
Pesaran ,
Naveen
1
Verma ,
and Sigurd
1
Wagner
1Electrical
Engineering, Princeton University, Engineering Quadrangle, Olden Street, Princeton, NJ 08544
2Center for Neural Science, New York University, 4 Washington Place, New York, NY 10003
3Integra LifeSciences Corporation, 103 Morgan Lane, Plainsboro, NJ 08536
Motivation
Why Polyimide?
Results
For eventual clinical applications of cortex-penetrating brain-machine
interfaces, neural stimulation and recordings must be performed over an
extended period of time.
Current penetrating arrays incorporating hard, planar backplanes have a large
mechanical mismatch with the soft, curvilinear tissue of the brain, leading to
strain and actual damage on the surrounding neurons. Focusing on matching
mechanical properties (ie Young’s Modulus) of tissue and device materials,
we use polyimide to construct our backplane.
Electrophysiological testing was performed with in vivo implantation into rat
subjects at the Center for Neural Science at NYU.
However, current penetrating interfaces are critically limited by reliability
concerns. Microelectrode arrays, for example, have been observed to exhibit
diminishing performance over time, leading to eventual failure, making them
ill-suited for chronic recording of neural signals.
This project aims to improve the quality of neural recordings for in vivo
microelectrode arrays by combining materials and technologies used in
flexible electronics with collagen, a material used extensively in the field of
regenerative medicine.
This Utah Intracortical Electrode Array has
been used extensively in acute neural
recording and stimulation experiments but is
believed to not provide reliable long-term
recording for chronic systems.
Collagen Matrix
Implantation of the microelectrode arrays requires an invasive surgery where
skull bone and dura mater are removed. This disrupts the homeostatic
processes regulated by the implant-affected tissue layers, which may lead to
increased neuronal death and dendritic loss within this critical region.
Biological tissues
Muscle.......................................................280,000Pa (280kPa)
Spinal cord....................................................89,000Pa (89kPa)
Brain.................................................................1,000Pa (1kPa)
Substrate materials
Silicon........................................200,000,000,000Pa (200GPa)
Polyimide........................................2,500,000,000Pa (2.5GPa)
Our Microelectrode Array
Our device consists of 3 main components: a flexible backplane made of
copper conductors on polyimide; tungsten electrodes, coated with Parylene-C;
and a collagen layer introduced over and around the electrodes.
Fabrication
To fabricate the device, first, an acrylic jig is used to hold the electrodes with
the tips facing downward. The collagen layer is introduced over the back of
the electrodes. After wetting the collagen, the backplane is threaded through
the back of the electrodes.
Neural signals were recorded as a function of time. While local field
potentials were recorded, we did not manage to record isolate units.
Conclusions
The backplane is attached to the collagen using non-conductive epoxy. Silver
conductive epoxy is used to bond the electrodes to the side of the backplane
not touching the collagen layer.
The collagen matrix is a substrate seen in many connective tissues, promoting
dural repair and re-growth of neurons. It has been successfully used in
posterior fossa duraplasty procedures, where it is applied as an onlay graft.
By incorporating a collagen matrix into the microelectrode array, it is hoped
that the health of the tissue surrounding the electrode implant will improve,
making it suitable for chronic neural recordings.
• The device seems to be able to record neural signals above noise.
• The signals across all four electrodes look very similar. Though not
unusual, it might be due to coupling between the leads which carry the
signals from the electrodes to the connector.
Future Work
Finally, a connector is connected to the bottom of the backplane using
soldering paste. The Omnetics connector attaches to a complimentary
connector employed by our collaborators at NYU, providing a means to read
the neural signals off of the device.
• Try and record isolate active cells
• Modify array design to eliminate any possible capacitive coupling between
leads
• Scale array to fit in larger animals
• Observe recording performance of device over an extended period of time.